> They just solved a Millennium Problem for fuck's sake.

Oh wow, they can solve formal proof test driven development. That's great! I too think it's cool. It isn't going to replace lawyers, doctors, nurses, tradespeople, and everyone else that drives the economy that isn't LLM output. The most successful YC startups have been DoorDash, Coinbase, and Airbnb; regulatory arb and a crypto gambling site. I do not disagree you can spit out more code faster, or potentially better code with an average dev versus more expensive bespoke code with a senior dev. More code does not translate to economic gains directly. People do not want slop entertainment. No LLMs are needed to build everything in flight for the global energy transition (batteries, solar, wind, geothermal, etc).

AI Hallucination Cases Tracker - https://www.damiencharlotin.com/hallucinations/ (2041 cases identified so far)

80% of OpenAI, Anthropic's enterprise revenues come from 1% of their customers - https://news.ycombinator.com/item?id=49613331 - September 2026 (My note: show me these customers and their wildly more profitable enterprises from this incredible inference use; Meta scrapped their "AI focused" layoffs, Salesforce had to hire software engineers back)

AI is everywhere at work now, but the time savings are smaller than the hype - https://censuseasy.com/blog/does-ai-save-time-at-work - August 13th, 2026

> More than half of American workers have now used artificial intelligence on the job, according to a new U.S. Census Bureau survey. But when the government asked those workers how much time AI actually saved them, the answer was smaller than the hype: for about 7 in 10, it was two hours a week or less, or nothing at all. [My note: 0-2 hours a week! For trillions in capex!]

The State of Generative AI Adoption in 2025 - https://www.stlouisfed.org/on-the-economy/2025/nov/state-gen... - November 13th, 2025

The Impact of Generative AI on Work Productivity - https://www.stlouisfed.org/on-the-economy/2025/feb/impact-ge... - February 27th, 2025

> The next figure illustrates how generative AI-driven time savings vary with usage across occupations. Time savings and overall usage are highly correlated. Workers in the computer and mathematics occupation used generative AI in nearly 12% of their work hours, and they reported this saved them 2.5% of work time. By contrast, workers in personal service occupations used this technology in only 1.3% of their work hours, and it saved them only 0.4% of work time. The slope of the dashed regression line is 0.17, indicating that a 10 percentage point increase in the share of time spent using generative AI is associated with a 1.7 percentage point increase in the time saved as a share of hours worked.

"Feel the hope!" Like a religion, people want to believe when the data says otherwise. I have been in many churches in my life, but the church of HN is by far the most committed to their faith. If I'm the one coping, where's the proof? It doesn't exist. It is not my job to believe without objective data and evidence supporting an opinion or assertion. "In God we trust. All others must bring data."

I like to cite those St. Louis Fed links too but for the opposite reason. With the caveat that it's based on survey data, it is one of the few data sources that shows the real of AI in the real world. And it is pretty astounding. It is only used for 6% of all work hours but it saves 2% of all work hours.

That is very low usage, but that is a 33% savings!

Now let's put that trillions in capex into perspective. About 55% of people use it for 6% of all work hours. Approximating a bit, that's basically 3% of all of the work hours by all of the workers in all of US, and as above, AI saves a 3rd of that time, so AI is already saving 1% of all work hours in the US!

(In case you think this Math is off, the same St. Louis Fed articles above have similar numbers, and corroborates with other national data sources as well as research studies, and suggests that the impact of GenAI may already be showing up in "national level statistics" to the tune of a 1.3% bump in national labor productivity! With only this shallow level of adoption! In just 3 years! It took the computer revolution about 2 decades to show up in economics data.)

According to the BLS, employers pay $12.3 trillion for that work. So a 1% savings on that is a $123 billion. An up front ~$2T investment over 2 - 4 years that is already saving $123 billion annually in the US alone is pretty damn good actually, and will only go up as usage improves.

This is not hope, this is data, and you had already brought it with you! ;-)